16:00:37 #startmeeting openstack_ansible_meeting 16:00:43 Meeting started Tue Apr 3 16:00:37 2018 UTC and is due to finish in 60 minutes. The chair is evrardjp. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:00:44 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 16:00:47 The meeting name has been set to 'openstack_ansible_meeting' 16:00:59 o/ 16:01:22 #topic What happened since last meeting 16:01:31 prometheanfire has a new spec that needs some eyes: https://review.openstack.org/556586 16:01:40 other ppl have open specs too 16:01:49 tonyb has EOLed newton roles. The newton release will be EOL soon. 16:01:53 o/ 16:02:00 please review https://review.openstack.org/#/c/558539/ for that 16:02:05 o/ 16:02:17 evrardjp reminded on the ML the etherpad for vancouver forum: https://etherpad.openstack.org/p/YVR-openstack-ansible-brainstorming 16:02:30 evrardjp said he will do the usual release at the end of the week for all branches, including newton, and bump master. But due to pbr/pip unconstrained, the release has been delayed to this Tuesday. 16:02:40 (Some are gone already, pike pending) 16:02:53 and last but not least 16:02:55 KeithMnemonic will do a demo of SUSE config processor and Ardana after the meeting of April, 3rd (5PM utc). https://global.gotomeeting.com/join/868941021 16:03:07 #topic bug triage 16:03:30 evrardjp: Hey might have just found a bug, should https://git.openstack.org/cgit/openstack/openstack-ansible/plain/doc/source/user/aio/scenario-table-gen.html be empty? 16:03:31 I guess a few bugs will be linked to the new version of pbr/pip 16:03:40 spotz: yes it's fine 16:03:46 ok:) 16:03:57 building the docs would generate it 16:04:07 let's head to our first bug 16:04:09 #link https://bugs.launchpad.net/openstack-ansible/+bug/1760574 16:04:12 Launchpad bug 1760574 in openstack-ansible "pip2 fails to install pkgs requiring python3 (dnspython3, mypy, etc)" [Undecided,New] 16:05:12 did we have a bug for the pip thing odyssey4me ? 16:05:33 evrardjp not to my knowledge 16:05:53 yeah so we can't mark this as duplicate 16:06:10 I'll mark this manually as fix commited 16:06:22 ok for everyone? Critical + Fix commited 16:06:33 ja 16:06:41 next 16:06:43 #link https://bugs.launchpad.net/openstack-ansible/+bug/1760219 16:06:44 Launchpad bug 1760219 in openstack-ansible "Issues with osa-ops mnaio service config with existing entries" [Undecided,New] - Assigned to Shannon Mitchell (shannon-mitchell) 16:07:04 Amy Marrich (spotz) proposed openstack/openstack-ansible master: [DOC] Update for Newton EOL https://review.openstack.org/551316 16:07:15 I saw a patch for that 16:07:29 I will just remind the user to mark this in the bug 16:07:36 to mark the bug in the patch* 16:07:41 * cloudnull is late, o/ 16:08:24 Jean-Philippe Evrard proposed openstack/openstack-ansible-ops master: Allow users to skip services to cut down on conflicts https://review.openstack.org/557999 16:08:31 ok typo corrected. 16:08:44 ok next 16:08:46 looks like https://review.openstack.org/#/c/557999/ is looking to solve that bug 16:08:46 #link https://bugs.launchpad.net/openstack-ansible/+bug/1759896 16:08:47 Launchpad bug 1759896 in openstack-ansible "Placing certificate on keystone containers does not work" [Undecided,New] 16:09:02 odyssey4me: yes there was a typo in the bug I just fixed it 16:09:13 you can review :) 16:09:19 let's go to next one 16:10:32 There is documentation 16:11:00 pointed to it in the bug 16:11:12 maybe the bug is asking for more documentation, or pointing to an issue 16:11:23 but I am not sure to understand the problem 16:12:56 maybe link to the docs and see if that sorts it out? :) i'm not familiar enough w/ what he's doing to understand where it's going wrong. 16:13:14 ok 16:13:17 yeah 16:13:19 let's see 16:13:23 moving on then! 16:13:28 #link https://bugs.launchpad.net/openstack-ansible/+bug/1759873 16:13:30 Launchpad bug 1759873 in openstack-ansible "[osa 16.0.9] gnocchi does not build anymore" [Undecided,New] 16:14:06 I guess that one is the PBR issue 16:14:14 odyssey4me: am I wrong? 16:14:19 agree 16:14:23 ok 16:15:27 Marked it as fix commited and critical 16:15:37 next 16:15:39 #link https://bugs.launchpad.net/openstack-ansible/+bug/1759784 16:15:41 Launchpad bug 1759784 in openstack-ansible "Gate jobs don't cover Glance image upload to non-NFS file store" [Undecided,New] 16:16:21 Given a server that only has 2 NICs, is it better to bond them for redundancy or separate them for isolation? 16:16:35 when we were discussing the other day, odyssey4me said the basic computekit test uses glance w/ file store, so this should have coverage in branches that have the computekit test 16:17:39 yes 16:17:43 this will be easier to implement with https://review.openstack.org/551791 16:17:54 the others have the integrated with swift but computekit could be with file 16:18:04 agreed with what odyssey4me said there 16:18:25 if anyone has extra cycles for it, we can work on this 16:19:26 let's classify this bug, shall we? 16:19:31 ofc it's confirmed 16:19:42 but now... is that a wishlist, or is that an issue 16:19:47 for me code without test is a bug 16:19:49 :p 16:20:09 considering we learned computekit has coverage for this i dont think there is a bug at all 16:20:19 Kevin Carter (cloudnull) proposed openstack/ansible-role-systemd_service master: Add a documentation list default https://review.openstack.org/558553 16:20:29 the computekit test was implemented after the commit that broke file backend 16:20:31 iirc 16:20:33 but because it would be hard to backport to coverage I'd say it's maybe wiser to mark it as wishlist 16:20:57 do we have computekit in P? 16:21:22 well, right now we use compute kit for opensuse/centos... so it is gate tested 16:21:25 we do not 16:21:31 not in pike correct 16:21:36 yeah but not in the branch asked there 16:21:43 oh yeah, not in p 16:21:54 I'd classify this as confirmed and wishlist 16:22:08 it's not urgent because we have coverage in upper branches 16:22:19 anyone willing to backport the test suite is welcomed 16:22:25 wfm 16:22:29 good 16:23:03 next 16:23:05 #link https://bugs.launchpad.net/openstack-ansible/+bug/1759721 16:23:07 Launchpad bug 1759721 in openstack-ansible "Prepare the deployment host in OpenStack-Ansible" [Undecided,New] 16:24:07 oh I see 16:24:10 ah https://rdoproject.org/repos/openstack-pike/rdo-release-pike.rpm 16:24:13 yeah 16:24:15 same for suse 16:24:19 w're going to have to add a substitution there 16:24:22 very hidden thing 16:24:33 I can pick that up if you like 16:24:41 Merged openstack/openstack-ansible-ops master: Allow users to skip services to cut down on conflicts https://review.openstack.org/557999 16:24:48 looks valid , low hanging fruit and medium? 16:24:59 odyssey4me: I'd rather see other contributors fixing it but fine for me :) 16:25:08 yep 16:25:25 ok next then 16:25:30 #link https://bugs.launchpad.net/openstack-ansible/+bug/1759323 16:25:31 Launchpad bug 1759323 in openstack-ansible "Script to check for requirements" [Undecided,New] 16:26:20 mmm 16:26:23 that seem hard 16:26:34 how can we assume what will be installed? 16:26:46 the only way I can see is if we build some kind of wizard tool 16:27:10 (which could leverage the bootstrap-hosts role) 16:27:31 what would you think? 16:27:40 i don't understand the ask, personally 16:27:56 it seems outside the scope to build the openstack architecture for the deployer 16:28:14 ah yes, I remember this 16:28:15 I guess the deployer forgot a group 16:28:23 basically it's an inventory check 16:28:44 if you're deploying nova, you need glance right? but we allow you to deploy an environment without it 16:29:07 similar that if you have compute, but no nova control plane hosts... we allow that, but it won't work 16:29:16 odyssey4me: so you'll have ara reports for all jobs now ? :D 16:29:24 this guy was missing glance and the nova control plane stuff 16:30:18 It was me, when i was checking over i didnt notice they were missing, niether did other people i had look at it 16:30:43 ok 16:30:55 it would be useful to have something that can tell you if something you listed is missing required parts 16:30:55 not sure how we can help there though 16:31:10 how can we say you need glance? 16:31:20 what if you had it with something else 16:31:29 maybe a pre-check like we have detected the followign componetns for install - keystone, haproxy, bla bla bla .. press Y to proceed 16:31:39 just 1 time in the start 16:31:53 yeah so building some kind of wizard 16:31:55 or something like install-summary.py that will just show what components/env is detected 16:31:58 i was thinking either a precheck or a seperate script for troubleshooting/checking configs 16:32:08 evrardjp sure, but it'd be nice to have some way of checking the config and outputting some warnings 16:32:09 just to check if the install script/inventory sees all 16:32:21 similar to " openstack-ansible setup-infrastructure.yml --syntax-check " 16:32:28 --setup-check 16:32:37 all it needs to do is output the projects it sees 16:32:46 odyssey4me: let's think about an actual deployer example which has things running in multiple dcs.. 16:32:52 (you know who) 16:32:52 Don't forget the case where someone set up some component a different way, so it isn't in the o-a yaml. 16:33:07 evrardjp this is to help the newbies, not to prevent complex configs 16:33:10 persia: exactly! 16:33:36 it helps people who're trying to get to grips with the basics 16:33:37 yes us newbies need help 16:33:37 ok so I think with this conversation we have a wishlist item 16:33:46 yes, absolutely 16:33:46 RandomTech: :) 16:33:53 wizard like 16:33:55 ok 16:34:16 got it 16:34:37 I don't think it needs a wizard, but we will need to add some checks which validate known dependencies for services that have multiple groups, and then also for inter-service deps. 16:35:05 could this be documented better in sample user configs, etc. ? 16:35:07 Perhaps that, to you, is a wizard. :) 16:35:19 perhaps sample configs aren't optimal 16:35:29 mattt: we have sample configs 16:35:44 last i looked they were super detailed though, kind of hard to read 16:36:00 mattt: the user stories should be easy to follow 16:36:10 i'm just wondering if there's another way to help avoid this situation 16:36:12 more than the same config 16:36:21 we could probably edit the samples to use anchors like the ceph example does - that'd make it a bit easier to read 16:36:34 mattt: i agree that they can be hard to read, the problem was that multiple people missed the fact that i only had part of the compute services defined 16:36:35 odyssey4me: I've seen so many ppl confused by that 16:37:22 well, short of saying - you need to understand openstack before you can put your config together (which is not friendly), I'm not sure that docs can really solve it unless we try to document our group deps and the openstack deps 16:37:27 RandomTech: I think if we clarify the groups in the future, it should help 16:37:37 i believe so 16:37:55 ok so that will be part of a spec I am writing for now 16:37:58 let's move on 16:38:03 #link https://bugs.launchpad.net/openstack-ansible/+bug/1758144 16:38:05 Launchpad bug 1758144 in openstack-ansible "resolv.conf in containers set too late" [Undecided,New] - Assigned to Logan V (loganv) 16:38:32 have not had time to work on that yet sadly 16:38:36 hopefully this week 16:38:37 logan-: I guess with the latest pip things you couldn't work on it? 16:38:43 ok 16:38:45 thanks 16:39:13 next 16:39:15 #link https://bugs.launchpad.net/openstack-ansible/+bug/1755821 16:39:16 Launchpad bug 1755821 in openstack-ansible "config_template fails to parse template if it contains a comment with leading spaces" [High,New] - Assigned to Kevin Carter (kevin-carter) 16:39:23 cloudnull: ? 16:39:51 * RossSchulman[m] didn't realize there was a bug triage meeting going on, apologies, will ask question again later. 16:40:05 RossSchulman[m]: no worries 16:41:32 cloudnull: doesn't seem here, so let's move on 16:41:38 #link https://bugs.launchpad.net/openstack-ansible/+bug/1752073 16:41:39 Launchpad bug 1752073 in openstack-ansible "Neutron firewall_v2 driver missing" [Undecided,New] - Assigned to Markos Chandras (hwoarang) 16:42:27 evrardjp: maybe this can be re-assigned to mbuil to work on it with peri 16:42:33 ok 16:44:01 next 16:44:08 #link https://bugs.launchpad.net/openstack-ansible/+bug/1743032 16:44:09 Launchpad bug 1743032 in openstack-ansible "Galera cluster maintenance in OpenStack-Ansible" [Undecided,New] - Assigned to Kevin Carter (kevin-carter) 16:45:21 hwoarang, evrardjp: fine with me :) 16:45:38 mbuil: thanks! 16:45:49 I've updated the bug on LP, added it to you 16:46:05 let's move to next bug 16:46:10 #link https://bugs.launchpad.net/openstack-ansible/+bug/1737827 16:46:11 Launchpad bug 1737827 in openstack-ansible "(ceph-client): setting 'nova_ceph_client' results in deployment where volumes can't be attached to VMs" [Undecided,New] - Assigned to Andy McCrae (andrew-mccrae) 16:46:17 andymccr: ? 16:46:31 Merged openstack/openstack-ansible-tests master: Do not generate the html ARA report any more https://review.openstack.org/557921 16:46:45 should we mark this as incomplete? 16:47:34 ok let's move on 16:47:41 #link https://bugs.launchpad.net/openstack-ansible/+bug/1729263 16:47:42 Launchpad bug 1729263 in openstack-ansible "Swift (master) transient tempest failures under centos " [Undecided,New] 16:48:15 I haven't seen that for a while, but I haven't worked on swift role testing for a certain time 16:49:31 Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible master: Ensure that RDO/SuSE series repo instructions stay up to date https://review.openstack.org/558558 16:49:38 swift is also a very heavy memory consumer 16:49:47 so this might be hidden 16:50:35 (the memory consumption is also why I think we have so many issues in the integrated gates, we start to spike in memory usage after the swift playbook, to what I have noticed) 16:50:58 are we still using 8 GB instances? 16:51:23 yes 16:51:41 do you think it's possible to have bigger instances in os infra? 16:51:46 it is not 16:52:13 ok 16:52:22 we could make swift use a multi-node deployment if we like, but that makes testing outside the gates harder 16:52:24 let's continue the tests refactor then 16:52:39 odyssey4me: not really if we justify the use of vagrant 16:52:55 Merged openstack/openstack-ansible-tests master: Remove unused tempest variables https://review.openstack.org/558524 16:53:07 well, not everyone has a laptop with enough ram to do multiple nodes 16:53:07 and only for some scenarios 16:53:21 that's true 16:53:27 I would rather we see if we could reduce that memory footprint a bit 16:53:49 maybe it's worth discussing with swift ppl 16:54:02 In the meantime let's move on to next bug 16:54:15 we have no more details for the bug in itself 16:54:22 I'll mark it as incomplete 16:54:42 the last one is the everlasting ceilometer one 16:54:44 #link https://bugs.launchpad.net/openstack-ansible/+bug/1719517 16:54:45 Launchpad bug 1719517 in openstack-ansible "Ceilometer policy.json includes python unicode encoding" [Undecided,New] 16:55:05 oh I guess it can be moved to fix released 16:55:13 wooot 16:55:25 FOLKS, it's been a while we got the chance to finish the triage! 16:55:29 congratulations! 16:55:37 /insert party gif here. 16:56:12 #topic open discussion 16:56:20 we have a few minutes for an open discussion if you want 16:56:30 so keep in mind we have a demo in 4 minutes 16:57:07 thanks KeithMnemonic 16:57:10 np 16:57:34 I will paste the url once again (even if that was on the first links of the meeting) 16:57:36 https://global.gotomeeting.com/join/868941021 16:57:58 anything else? 16:58:35 whats the demo about? 16:58:56 demoing the SUSE project Ardanda 16:59:05 which is the rebranded HPE ansible stuff 16:59:30 it's a way to learn of each other -- trying ways to collaborate 16:59:33 that was never collborated with you guys 16:59:34 anyway 16:59:41 of ya something i just noticed while using OSA for ceph 16:59:54 RandomTech: ? 16:59:57 10 sec 16:59:59 it only added osd from 3/5 hosts i listed 16:59:59 it is now opensource 17:00:17 not sure if this is isolated to me or if multiple people have ran into this 17:00:19 RandomTech: I guess it can be discussed outside the meeting 17:00:25 or file a bug :) 17:00:28 Ok wrapping up 17:00:34 thanks for your time everyone! 17:00:36 #endmeeting